{"version":1,"type":"story","url":"https://digestai.news/story/oracle-adds-rdma-and-multiplanar-failover-to-acceleron-for-ai-workload","json":"https://digestai.news/story/oracle-adds-rdma-and-multiplanar-failover-to-acceleron-for-ai-workload.json","markdown":"https://digestai.news/story/oracle-adds-rdma-and-multiplanar-failover-to-acceleron-for-ai-workload.md","slug":"oracle-adds-rdma-and-multiplanar-failover-to-acceleron-for-ai-workload","headline":"Oracle adds RDMA and multiplanar failover to Acceleron for AI workloads","summary":"Oracle unveiled new features for its Acceleron cloud networking suite at Oracle AI World on October 27, 2026. The updates include front-end RDMA connectivity, expanded NVMe block-storage options with up to 3x performance gains, and the Oracle Collective Communications Library (OCCL) for distributed AI workloads. These changes aim to improve data movement, reduce congestion, and enhance resilience in AI training and inference tasks by rerouting traffic and recovering from failures automatically.\n\nThe new capabilities address key pain points in AI infrastructure. Front-end RDMA reduces latency by moving data directly between machines, while OCCL dynamically balances traffic across multiple paths to avoid bottlenecks. Oracle claims these improvements help customers scale AI workloads securely without manual intervention. The company also highlights that existing applications can use the new features without modification, as front-end RDMA supports all queue pair types and verbs.","keyPoints":["Oracle adds front-end RDMA connectivity to Acceleron for faster AI data movement, reducing CPU overhead","Expanded NVMe block-storage options improve performance up to 3x for AI and database workloads","OCCL library enables dynamic traffic balancing and failover for distributed AI training without restarts"],"whyItMatters":"These updates make Oracle’s cloud infrastructure more competitive for AI workloads by cutting latency and improving fault tolerance, which could attract hyperscale customers running large-scale training and inference jobs.","category":{"slug":"hardware","name":"Hardware & Compute","url":"https://digestai.news/category/hardware"},"entities":{"companies":["Oracle","Oracle Cloud Infrastructure"],"models":[],"people":["Pradeep Vincent","Jag Brar","Joy Ganguly"]},"firstPublishedAt":"2026-10-27T17:05:02Z","updatedAt":"2026-10-27T17:05:02Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"Unite.AI","title":"Oracle Acceleron Adds Front-End RDMA and Multiplanar Failover for AI","url":"https://unite.ai/oracle-acceleron-adds-front-end-rdma-and-multiplanar-failover-for-ai","publishedAt":"2026-10-27T17:05:02Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Oracle adds RDMA and multiplanar failover to Acceleron for AI workloads\", 27 October 2026, https://digestai.news/story/oracle-adds-rdma-and-multiplanar-failover-to-acceleron-for-ai-workload","publisher":"Digest AI","title":"Oracle adds RDMA and multiplanar failover to Acceleron for AI workloads","datePublished":"2026-10-27T17:05:02Z","url":"https://digestai.news/story/oracle-adds-rdma-and-multiplanar-failover-to-acceleron-for-ai-workload"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}